Home Assistant runs into recovery mode with error "HTTP option 'server_host'. Watchdog feature disabled"

Hi All
Please help me to solve this issue.

My Home Assistant runs into the recovery mode after upgrade Home Assistant Core 2024.7.0. The full logs is show about server_host configure.

Home Assistant Core
2024-07-04 09:36:52.918 WARNING (MainThread) [homeassistant.components.hassio.handler] Found incompatible HTTP option 'server_host'. Watchdog feature disabled

I searched for a solution to fix it. which found that many suggestions to remove the “server_host” under the “http:” section, but there are no “server_host” in my configuration. So, I just remove all “http:” section. Then restart the Home Assistant , but the system still runs into recovery mode and report the same error log.

Anyone have an ideas for fixing it. Or in my step I took were wrong.

Thank you.

I had the same thing happen to me.
in addition to removing the entire http section from the configuration.yaml file, I also tried stopping several addons at the start as well as nodered but nothing.

I saw through ssh the server_host directive is contained in the config/.storage/http path.

I tried to remove it or let the system to generate a new http file but nothing. This directive is automatically added.

{
  "version": 1,
  "minor_version": 1,
  "key": "http",
  "data": {
    "cors_allowed_origins": [
      "https://cast.home-assistant.io"
    ],
    "server_host": [
      "0.0.0.0",
      "::"
    ],
    "ip_ban_enabled": true,
    "ssl_profile": "modern",
    "server_port": 8123,
    "login_attempts_threshold": -1,
    "use_x_frame_options": true
  }
}

I encountered the same error. Rolled back to backup.

Same error. +1

Same error for me too, no server_host parameter exists in my config.

I resolved my installation by removing the “legacy_api_password” authentication method as this has been removed from HA in this release. (Comment it out/delete it from the yaml config file.

I’ve now successfully upgraded.

5 Likes

Work! Thanks!

it works for me. Thank you.

I had to remove the ‘trusted_networks’ node from configuration.yaml to ‘fix’ the error.

Yup that fixed it

Thank you for this, was banging my head as to what was causing my failure after upgrading to 2024.7. I’m now back in business.

good day (daytime greeting)
I have started to have the same problem since I upgraded to the version in the image below, and I have been using this thread to troubleshoot the problem, but I cannot solve it.
The "legacy_api_password " you are referring to here
I am looking for the file path of "legacy_api_password ", but I can’t find it, so could you please help me?

I had the same issue today but I had no http section in my configuration.yaml. When using HA command line I did a “core check” and it errored with some missing files. My config was including scene.yaml and scripts.yaml but they were missing. Fixing this solved the issue.

Odd thing was I have 2 identical HA’s but only one errored. It also doesn’t help that the error log in HA was reporting wrong or incomplete data. Was only by accident I did “core check” and found the real issue.

I hade the same issue. For me it was the line !include knx.yaml which I added a few days ago. It seems that it blocked the other config files to load which caused the error?

I had the same problem this morning. I looked in the /homeassistant/home-assistant.log.1 file and it showed an error toward the beginning where one of my custom themes had an issue with an undefined color. I edited the themexxx.yaml correcting my error and rebooted. Problem solved.

I found that my configuration file has “Op” on line 8.

I’m not sure why it was there or what it does, but when I commented it out I could boot again normally.

Anyone else seen this before?

my home assistant went into recovery mode when applying 2024.9.3
It turns out that I was missing the word themes from the directive below
before

frontend:
  themes: !include_dir_merge_named*

after

frontend:
  themes: !include_dir_merge_named themes

Not sure how it did not complained before now :slight_smile:
so it looks like that now

Thanks, works for me.

I had the same error today, when upgrading from 2024.10.1 to 2024.10.2.

Based on information in this thread, I tried simply moving the file .storage/http aside, and restarting HA. That cleared the problem. But I was curious, and I tried copying the file back into place, but without the “server_host” part. That also worked, after a restart. So, still curious, I copied the file back into place again, so that it was identical with what I’d started with. And again, a system restart (of HA, not the whole RPi) was again successful. No error. Diffing the http file with the copy I had set aside originally, some lines had changed ordering, so it seems that it was rewritten during the reboot. So one more experiment: I removed the http file entirely, and again rebooted the whole system. The http file was regenerated, complete with the problematic 'server_host" line.

So, what’s going on here? My problem has been solved, but it’s hard for me to believe that I won’t see it again someday, since the “server_host” parameter is still in the config.

I just ran into the same issue when doing an upgrade and working in some new template sensor in parallel. I thought I had the sensor definition correct and did the restart after updating the HA core.

I got the error message and looking through my configuration.yaml I realized that there was a mistake preventing it from proper execution. After fixing that HA booted without any issue and the above error was gone as well

Seems like that error is thrown somewhat randomly, when other issues are present.